From the Princeton Weekly Bulletin, May 26, 1997
How fast does universe expand?
By Jacquelyn Savani
Astrophysics Professor Edwin Turner, his former student Tomislav Kundic (a post-doctoral fellow at California Institute of Technology) and Princeton graduate student Wesley Colley headed a team that has worked out a new method for determining the expansion rate of the universe. Measuring that rate--Hubble's constant--has long been of keen interest to astrophysicists because it is a key parameter for understanding the universe--its age, history, structure and ultimate fate.
The method uses gravitational lensing to establish a distance scale from which Hubble's constant can be computed.
Gravitational lensing occurs when we look out at a distant source of radiation in the universe (in this case a quasar), and somewhere along the line of sight between us and the quasar exists a massive object (in this case a galaxy and an associated cluster of galaxies). The gravitational field of the intervening object bends light rays in the same sense that a lens bends light. So we see light from the quasar as affected by the gravitational lens. In this case we see two images of the same quasar, identical in appearance but slightly separated in the sky, a sort of cosmic mirage.
But the light from one image of the quasar reaches us before the light from the other image, because the space along which the two sets of light rays travel is differently curved. One image traverses more miles than the other. So when the quasar itself fluctuates in brightness, millennia later we see that change in one image and then days later the same change in the other image.
The big question has been how many days later. Attempts to answer that question for one gravitational lens system, called 0957+561, have ranged astrophysicists into two camps: proponents of the short delay (just over 400 days) and proponents of the long delay (around 540 days).
Prediction verified
Turner's team has measured the delay by first publishing a prediction and then verifying it by observation.
The team started observing the gravitational lens system 0957+561 late in 1994. Soon afterwards they observed a strong and sharp fall in brightness in one of the quasar's images.
"By June of '95, the end of the first observing season, we realized that we could predict from image `A' what image `B' would be doing," says Turner. The team published their prediction in a paper, "An Event in the Light Curve of 0957+561A and Prediction of the 1996 Image B Light Curve," which appeared in Astrophysical Journal Letters in December 1995. Kundic is the paper's first author, and the prediction is part of his Princeton PhD dissertation supervised by Turner.
"We saw a fall in brightness of the leading image," says Turner, "and published a prediction that said if the short delay is right, the second quasar image will fluctuate in brightness in February of 1996; if the long delay is right, then the fluctuation will occur in June." It happened in February.
"This repetition of the event gave us the delay very precisely," said Turner, "better than one percent accuracy."
Finding the distance
But how does one go from the delay to the distance scale and from there to the expansion rate of the universe?
A shorthand way of describing the calculation, Turner says, "is that the delay multiplied by the speed of light gives the difference between the distances to the two images of the quasar. And the rest is high school geometry. If you know a lot of angles and a distance, you can work out a lot of other distances.
"Once you have the distance," he goes on, "the next step is to divide the distance to the object by its recessional velocity in order to get the expansion rate of the universe." How fast the object is moving away from us on earth--its recessional velocity--is indicated by how far its spectrum is shifted to the red. Measuring the recessional velocity or determining redshift is the easy part. It's getting the distance to divide by the redshift that has proved the stumbling block.
Astronomers have been working for decades to get a distance for the determination of Hubble's constant by using, among other things, a kind of star called a Cepheid variable. That method, dubbed the "distance ladder" by astronomers because of its numerous steps, would require a whole course, said Turner, to explain to a physics undergraduate major. The gravita-tional lens method would take one lecture.
Hubble's constant: 64
According to the Turner team calculations, the universe is expanding at the rate of 64 kilometers per second per megaparsec.
If there had been no matter in the universe to slow down expansion, the universe would be 15.3 billion years old, "which means," says Turner, "since there is stuff, the universe is younger than 15.3 billion years old." And if there is enough stuff to close the universe--i.e., to halt its expansion--then the universe with a value of 64 for Hubble's constant would be 10.2 billion years old. Turner shakes his head at the 10.2 billion: "That's a universe too young to contain the oldest stars," he says.
If the universe is older than 10.2 billion years, that means there isn't enough matter to close the universe and it will go on expanding. But Inflation, today's leading theory of how the universe underwent colossal expansion after a hot Big Bang, seems to depend on there being enough matter to close the universe, a value which cosmologists refer to as "omega = 1."
The next step, according to Turner, is to repeat the measurements using other gravitational lens systems. And his team is proceeding with a research program to do just that.
Using gravitational lensing to derive a distance scale and Hubble's constant is not a new idea. It was first advanced in the 1960s by a graduate student in Scandinavia, Sjer Refsdal, in his PhD thesis long before the first gravitational lens was discovered in 1979 (that lens system is 0957, the same one used by the Turner team). Refsdal's idea was so far ahead of its time that his thesis committee rejected the thesis as too speculative. He had to add more conventional parts to get the thesis accepted for his degree. Yet that speculative part has turned out to be what the Hamburg professor is most famous for today.
Quick change, remote viewing
For a few years after the first gravi-tational lens system was discovered, astronomers tried to pin down the delay between the two images in order to get a distance scale and Hubble's constant. Their efforts amounted to little because of the way observatories function.
Typically an observer is assigned a block of nights to use a telescope. But to measure the delay, what is required is half an hour of telescope time night after night for months. And for each of those half-hours, the same instrument has to be affixed to the telescope to record the brightness of the images. Conventional observatories are designed so that changes in the instru-ments used in conjunction with a telescope (a camera or spectrograph, for instance) take time--too much to accommodate an observing project that requires a little time with the same instrument night after night.
So it took a new kind of observatory for Turner's team to carry on the observing project that would yield the delay. The Apache Point Observatory in New Mexico with its 3.5 meter telescope was built by a consortium of institutions, including Princeton. The facility was designed for quick instrument change and for remote viewing--two technological innova-tions essential to the success of the Turner team project.
Remote viewing means that the telescope can be operated through the internet from a computer located anywhere. So night after night, Turner and graduate students took turns coming to the Peyton Hall basement on the Princeton campus to operate the telescope in New Mexico.
Turner, who is director of Apache Point's 3.5 meter telescope, says that as soon as that new kind of observatory was proposed in the early '80s, he began to envision and to plan the observing project that he and his students carried out some 15 years later.
The paper that reports the delay and the expansion rate of the universe, "A Robust Determination of the Time Delay in 0957+561A,B and a Measure-ment of the Global Value of Hubble's Constant," will appear in the June 10 issue of Astrophysical Journal. Along with Turner, Kundic and Colley, the paper's authors include Richard Gott, James Rhoads and Yun Wang of Princeton; Louis Bergeron, Karen Gloria and Daniel Long of Apache Point Obser-vatory; Sangeeta Malhotra of Caltech; and Joachim Wambsganss of Astro-physikalisches Institut Potsdam.